#1
9.0 Weirdness
OpenAI internal model autonomously disproves 80-year-old Erdos unit distance conjecture
An OpenAI model just killed a math problem Erdos posed in 1946 in one shot, and human mathematicians say the proof is elegant and real.
AI moves from tool to autonomous discoverer, forging unexpected connections between geometry and algebraic number theory; changes institutions of mathematical proof and what "original research" means.
What Happened
The source trail is attached to the daily scan.
Caveat
Needs careful wording around "autonomous"; cite OpenAI and mathematician validation rather than implying peer-review finality.
Source Trail
- No public source URL captured yet.
#2
9.0 Weirdness
Figure AI's viral livestream of F.03 humanoid robots sorting packages
Millions can't stop watching humanoid robots sort packages like prestige TV, while the job quietly turns into training data.
Humans are mesmerized by robots doing drudgery as entertainment while the same behavior signals labor replacement and data capture.
What Happened
The source trail is attached to the daily scan.
Caveat
Treat autonomy as Figure's claim unless independent validation appears.
#3
8.5 Weirdness
Indian factory workers wearing head-mounted cameras to train humanoid robots
Factory workers are strapping cameras to their heads to teach robots their every move, then realizing the robots are the point.
Workers are paid to record the gestures and tacit knowledge that may automate them; "training your successor" as labor horror, not abstract AI doom.
What Happened
The source trail is attached to the daily scan.
Caveat
Stronger with direct Scroll details and avoiding speculation beyond the reported data pipeline.
Source Trail
- No public source URL captured yet.
#4
8.0 Weirdness
Anthropic reportedly paying SpaceX $1.25B/month for Colossus compute
Anthropic may be nearing profitability while paying SpaceX $1.25 billion a month for compute. The cloud is now a rocket-company line item.
AI scaling economics now look like planetary infrastructure finance, with AI labs renting supercomputers from a rocket company.
What Happened
The source trail is attached to the daily scan.
Caveat
Keep it distinct from prior cloud infrastructure coverage by centering the financial weirdness and filing artifact.
Source Trail
- No public source URL captured yet.
#5
7.5 Weirdness
GitHub breach via malicious VS Code extension exposes roughly 3,800 internal repos
One poisoned VS Code extension reached roughly 3,800 internal GitHub repos. The tool you trust to write code is now the blast radius.
The trusted builder interface becomes the attack surface, especially as developers install more AI/coding extensions and grant them broad local access.
What Happened
The source trail is attached to the daily scan.
Caveat
Internal repos, no evidence of customer data outside affected repos per reporting; don't overstate.
Source Trail
- No public source URL captured yet.